DESCRIPTION: Only a tiny % of tumor cells within the tumor tissue is motile and capable of escaping and initiating metastasis. The gene signature of these motile tumor-initiating cells predicts clinical outcome in breast cancer patients. Patients exhibit heterogeneous responses to drugs targeting the cytoskeleton, which is a common target for therapies aiming to inhibit tumor cell growth, motility and invasion. To date, there is no clinical assay to 1) isolate the highly motile cell subpopulations from patient biopsies, and 2) investigate their responses to cytoskeleton-targeting drugs. A clinically relevant method for identifying optimal individualized therapies is to screen the effects of drugs on patient-derived tumor specimens transplanted in mice (tumorgrafts). However, this method suffers from four major drawbacks: 1) only 28-37% of patient-derived tumorgrafts grow in mice, 2) this method, if successful, requires 4-12 months, 3) it requires large tissue specimens, and 4) it examines the effects of drugs on the bulk tumor cell population, and not specifically on the motile tumor-initiating cells. It is thus urgent 1) to deveop a clinically relevant technology that will enable physicians to accurately and rapidly assess the relative abundance of motile cell subpopulations in primary tumors at the time of initial diagnosis, and 2) to understand how cytoskeletal-targeting drugs affect the properties and function of these motile cells within a heterogeneous tumor in order to develop personalized therapies. Such technology should be high- throughput, rapid with superior detection sensitivity, and capable of studying the responses of motile single cells isolated from a small number of tumor cells harvested from patient biopsies. The microfluidic technology meets all these requirements. In Aim 1, we will develop, test and validate a microfluidic assay that predicts the metastatic propensity of human breast cancer cell lines and patient-derived breast cancer cells. Because of the heterogeneous responses of cancer patients to cytoskeleton-targeting drugs, we will use our microfluidic assay to rapidly assess the responses of 8 human breast cancer tumorgrafts as well as patient-derived breast cancer cells to different cytoskeletal modulators to identify optimal personalized therapies (Aim 2). Most of the drugs to be tested are FDA-approved or in clinical trials, ensuring that this study could rapidly impact the clinical treatmen of breast cancer. Since cell properties regulate migration, invasion and metastasis, we will use a novel high-throughput, single-cell imaging technology to define the morphological, and molecular signatures of migratory vs. non-migratory breast cancer cells isolated by the ?-fluidic assay (Aim 3). In view of compelling data showing the involvement of giant obscurins in cancer metastasis, we will evaluate giant obscurins as a novel biomarker for breast cancer (Aim 4). We will also examine how the expression profile of giant obscurins: 1) is altered in migratory versus non-migratory cells; 2) regulates the responses of breast cancer cells to cytoskeletal drugs; 3) affects the morphological and molecular signatures of breast cancer cells; and 4) how ectopic expression of an obscurin signaling cassette can suppress tumorigenicity and metastasis.